Perfluorinated Compounds (PFCs)

PFCs and diabetes/obesity
On this page:

Perfluorooctanoic acid (PFOA) and perfluorooctanesulfonic acid (PFOS) have been used in the manufacturing of Teflon, Gore-Tex, and Scotchguard. They are perfluorinated compound (PFCs), also known as perfluoroalkyl substances (PFAS). Low levels of exposure are ubiquitous in the blood of essentially all residents of industrialized countries, via food, drinking water, house dust, and air (Steenland et al. 2013).

While initially assumed to be inert and non-toxic, these substances are now thought to have the ability to affect the immune system (and carcinogenicity) at exposure levels found in the general population. Existing drinking water limits may be 100-fold too high (Grandjean and Clapp 2015). Note that industry studies showing immunotoxicity were conducted decades ago, but not released to the public until recently-- long after PFCs contaminated groundwater systems throughout the world (Grandjean 2018). PFCs are a type of persistent organic pollutants.

Type 1 Diabetes

A small study from Italy found that PFOS levels were higher in children and adolescents with new-onset type 1 diabetes than in controls without diabetes. PFOA levels were not associated (Predieri et al. 2015). A study from Denmark found that PFOA levels in childhood were associated with lower beta cell function in adolescence (Domazet et al. 2016).

In an area with high PFC exposure levels (see Mid-Ohio Valley Studies section below), a study of community members looked for autoimmune diseases, including type 1 diabetes. They did find an association between the autoimmune disease ulcerative colitis, but not type 1 diabetes (Steenland et al. 2013). A furthers study also found no increased risk of type 1 diabetes with PFC exposure-- in fact there was a decreased risk (Conway et al. 2016). PFCs are potentially immunotoxic, which could favor the development of autoimmune diseases (or suppress the immune system) (Corsini et al. 2014).

A laboratory study on non-obese diabetic mice (NOD mice), a model of type 1 diabetes, found that exposure during development to perfluoroundecanoic acid, PFUnDA increased the development of insulitis, which occurs before diabetes development in these mice. It also increased cell death in the islets. However, it did not lead to accelerated diabetes development overall (Bodin et al. 2016). A mixture of persistent organic pollutants (that included PFCs) led to early signs of autoimmunity development in NOD mice (Berntsen et al. 2018).

A report on immunotoxicity of PFOA and PFOS by the National Toxicology Program (NTP) finds that, "PFOA is presumed to be an immune hazard to humans based on a high level of evidence that PFOA suppressed the antibody response from animal studies and a moderate level of evidence from studies in humans. Although the strongest evidence for an effect of PFOA on the immune system is for suppression of the antibody response, there is additional, although weaker, evidence that is primarily from epidemiological studies that PFOA reduced infectious disease resistance, increased hypersensitivity-related outcomes, and increased autoimmune disease incidence" (NTP 2016). The report also finds that PFOS "is presumed to be an immune hazard to humans."

When exposed to intestinal cells in a laboratory, a mixture of PFOS and other chemical surfactants caused leakage through the tight junctions (although PFOS seemed to protect from this effect) (Glynn et al. 2017); a leaky gut is linked to type 1 diabetes (see the Diet and the Gut page). Prenatal exposure to PFCs is also associated with infectious disease in early life (Goudarzi et al. 2017); infections are linked to type 1 diabetes (see the Viruses and Bacteria page). PFOS levels were associated with lower vitamin D levels in the U.S. population, especially in whites and the elderly, while PFHxS levels were associated with higher vitamin D levels (Etzel et al. 2018); low vitamin D levels are linked to type 1 diabetes (see the Vitamin D Deficiency page).  

Type 2 Diabetes, Body Weight, and Metabolic Syndrome

Lower- Level Exposures in Humans

Longitudinal Studies in Humans

The strongest evidence for the ability for environmental exposures to contribute to the development of diabetes comes from longitudinal studies. These are studies that take place over a period of time, where the exposure is measured before the disease develops.

Background exposures to PFOS and PFOA in the late 1990s were associated with higher type 2 diabetes risk during the following years in a prospective case-control study of women from the U.S.-based Nurses' Health Study II (Sun et al. 2018).

A long-term study of French women estimated exposure to PFCs based on French food contamination data and dietary consumption data. It found an association between type 2 diabetes and PFOA, especially in non-obese women, and an association between type 2 and PFOS in non-obese women (Mancini et al. 2018).

Overweight and obese adults with low level exposures (from Boston, MA and Baton Rouge, LA) with higher PFC levels had greater weight regain after a diet, especially women (Liu et al. 2018). Also in U.S. residents, a study found that PFOS and PFOA were associated with insulin resistance, beta cell function, and HbA1c. After 4.6 years of follow-up, however, these chemicals did not appear to affect the incidence of diabetes or changes in these markers (Cardenas et al. 2017).

A long-term study from Korea found that there were differences in PFHxS and PFDoDA levels between participants with and without diabetes, and while PFCs were not associated with BMI, they were associated with higher total cholesterol, LDL ("bad") cholesterol, and triglycerides and with lower HDL ("good") cholesterol (Seo et al. 2018).

Cross-Sectional Studies in Adults

Cross-sectional studies are studies that measure exposure and disease at one point in time. These provide weaker evidence than longitudinal studies, since the disease may potentially affect the exposure, and not vice versa.

A cross-sectional study of elderly Swedes found that the PFCs perfluorononanoic acid (PFNA) and PFOA were significantly related to diabetes in a non-linear manner. PFOA was also related to insulin secretion, but none of the PFCs were associated with insulin resistance. The exposures encountered in this study were typical of the general population (Lind et al. 2014). These authors also found that various PFCs were associated with various measures of metabolism, suggesting that each PFC may have different effects (Salihovic et al. 2018).

A cross-sectional study of Canadian adults found associations between some PFCs and cholesterol levels, but not glucose levels or metabolic syndrome (Fisher et al. 2013). 

A cross-sectional study of Americans also found associations between some PFCs and cholesterol levels, but not insulin resistance or body size (Nelson et al. 2010). (It seems like associations between PFCs and higher cholesterol levels are pretty consistent across studies, especially in studies of more highly exposed people, but also among less exposed, e.g., Eriksen et al. 2013). Despite using in part some of the same dataset as Nelson et al., (NHANES), Lin et al. (2009) found links between PFCs and various measures of blood glucose in Americans. For example, they found that in adolescents, higher PFNA levels were associated with higher blood sugar levels and cholesterol levels. In adults, higher PFNA levels were associated with higher beta cell function, and higher PFOS levels were associated with higher insulin levels, higher beta cell function, and insulin resistance. Another study using NHANES data found that different types (isomers) of PFOA were variously associated with higher blood glucose levels (and others with lower HbA1c), higher beta cell function, higher HDL and total cholesterol. Various types of PFOS were associated with higher beta cell function, lower HDL, and lower triglycerides. Both PFOS and PFOA were indicators of metabolic syndrome (Liu et al. 2018). And, also with NHANES data, PFOA (but not other PFC) levels were associated with diabetes in men (not women) and with total cholesterol in adults (He et al. 2018). Another analysis of NHANES data, looking at multiple years, found that PFNA was associated with increased risk of metabolic syndrome and well as several individual components, while the highest levels of PFHxS were associated with elevated triglycerides. Other PFCs were associated with decreased risk of at least one outcome (Christensen et al. 2018). Both obesity and gender affect the relationship between PFCs and cholesterol levels in NHANES: in obese males (but not in non-obese males), there were positive associations between total and LDL cholesterol with PFOA and PFNA. In obese females, total cholesterol levels increased in tandem with levels of PFDA, PFNA, and Me-PFOSAA, and there was a positive association of LDL cholesterol with PFOS, PFDA, and PFNA (Jain and Ducatman, 2018).

A study of working-aged Taiwanese adults found that those with higher PFOS levels had a higher risk of impaired glucose homeostasis and diabetes. However, those with PFOA, PFNA, and PFUA had a lower risk (Su et al. 2016). A study of Chinese adult men found that PFC levels were associated with various markers of metabolism (Wang et al. 2017). Another study of Chinese adult men found that PFC levels were associated with metabolic syndrome (Yang et al. 2018).

A trial of elderly adults from Korea found that while PFC levels were associated with insulin resistance, supplementation with vitamin C reversed these effects (Kim et al. 2016).

Cross-Sectional Studies in Children

A cross-sectional study of Danish children found that in overweight children, higher PFC levels were associated with higher insulin levels, higher beta cell activity, higher insulin resistance, and higher triglycerides. There was no association between these and PFCs in normal-weight children (Timmermann et al. 2014). A study of obese U.S. children found that PFC levels were associated with various metabolic measurements, including cholesterol levels, but not blood glucose levels (Khalil et al. 2018). Also in U.S. children, levels of PFOA and PFNA were associated with total cholesterol levels (Jain and Ducatman 2018).

A study of 2 year old Korean children found that those with higher levels of various PFCs were shorter, and those with higher levels of PFNA weighed less than those with lower exposures (Lee et al 2018). Also in Korean children, PFUnDA levels were associated with higher total and LDL cholesterol levels (Kang et al. 2018).
Factsheet on PFCs from the National Institute of Environmental Health Sciences (NIEHS)

Exposures During Development

Evidence is growing that exposure to pollution during critical developmental periods, such as in utero or during childhood, may have effects later in life.

A meta-analysis of 10 studies found that early life exposure to PFOA was associated with a higher risk of being overweight in childhood, and a higher BMI in childhood (Liu et al. 2018).

In a study from Denmark, in utero exposure to PFOA was associated with higher weight, overweight/obesity, higher waist circumference, and higher insulin levels in female offspring at age 20 (with similar results in males but fewer data points). Other PFCs, including PFOS and PFNA did not show any associations (Halldorsson et al. 2012).

In British girls, exposure to PFOS in the womb is associated with lower birth weight, but then higher weight at age 20 months (Maisonet et al. 2012). Other studies have also found associations between PFCs and birth weight as well (Apelberg et al. 2007; de Cock et al. 2016Fei et al. 2007; Lenters et al. 2016; Minatoya et al. 2017; Rokoff et al. 2018; Shoaff et al. 2018Starling et al. 2017Washino et al. 2009; Woods et al. 2017). Further study of the British girls at age 9 found that prenatal PFOA and PFOS levels were associated with percent body fat in different ways, depending on the mother's level of education. PFHxS and PFNA were not associated (Hartman et al. 2017). One also found associations between PFCs and levels of adiponectin, a hormone that plays a role in glucose levels (Minatoya et al. 2017). Another did not find statistically significant associations between prenatal PFC levels and adiponectin or leptin, another energy-related hormone, although the levels of both hormones were higher in infants with higher PFOA levels (Buck et al. 2018).  

A study from Sweden and Norway found positive associations between maternal PFOA and PFOS levels and child overweight/obesity at 5 years of age (Lauritzen et al. 2018).

A Danish study, however, did not find an association between PFOA or PFOS levels in mothers during pregnancy, and body mass index, waist circumference, or risk of overweight in their children at 7 years of age (if anything, the more highly exposed children were thinner than the others, although the difference was not statistically significant) (Andersen et al. 2013). These same authors had also found that maternal PFC levels were associated with lower body weight in the first year of life (Andersen et al. 2010). A different study from Denmark found that childhood levels of PFOS was associated with higher waist circumference at ages 15 and 21. PFOA levels in childhood were associated with lower beta cell function in adolescence (Domazet et al. 2016).

A study from Boston, Massachusetts found that PFC levels in mothers were not associated with metabolic changes in children, although children with higher PFAS levels had lower insulin resistance (Fleisch et al. 2017). However, another study of the same cohort by the same authors found that maternal PFAS levels were associated with small increases in weight-related measurements in girls in mid-childhood (Mora et al. 2017). And, these authors also found that levels of various PFCs were associated with changes in various cholesterol measurements, and not all the changes were detrimental (Mora et al. 2018). A separate study from Cincinnati, Ohio found that higher PFOA levels in the womb were associated with a lower BMI up to age 2 (Shoaff et al. 2018).

A Spanish study found that prenatal PFC levels were not generally associated with a higher BMI and other measures of metabolism through age 7 (Manzano-Salgado et al. 2017).

A study from Greenland and the Ukraine found that neither PFOA nor PFOS levels in mothers during pregnancy were associated with their children being overweight at ages 5-9. However, the children did have higher waist-to-height ratios (Høyer et al. 2015). A study from the Faroe Islands found that maternal PFOS and PFOA (but not PFHxS, PFNA or PFDA) levels after childbirth were associated with higher BMI in the offspring at 18 months and 5 years of age (Karlsen et al. 2017).

A Japanese study found that PFOS levels were associated with reduced fatty acid levels in pregnant women. These polyunsaturated fatty acids are essential for fetal growth. The female babies also had a lower birth weight if exposed to higher levels of PFOS (these associations were not found with PFOA or in male babies) (Kishi et al. 2015). A Taiwanese study also found that PFC levels were associated with lower birth weight in girls. Levels were not associated with weight through age 11 in either sex, but were associated with lower height (Wang et al. 2016). A Canadian study found that maternal PFOA levels were associated with a lower birth weight (Ashley-Martin et al. 2017).

In China, umbilical cord levels of PFASs were associated with various measures of growth at birth and at 19 months. Associations varied by specific PFAS and by sex (Cao et al. 2018).

Mid-Ohio Valley Studies: Higher-Level Exposures in Humans

An area of the Mid-Ohio Valley has been contaminated by high levels of PFOA. Research suggests an increased risk of mortality due to type 2 diabetes in workers occupationally exposed to PFOA, as compared to other DuPont workers (Steenland and Woskie, 2012). Previous research suggest an association between diabetes and PFOA exposure in workers as well (Lundin et al. 2009). Workers in the Mid-Ohio Valley were exposed to PFOA in a chemical plant that produced Teflon. Emissions from this plant polluted the drinking water of the nearby community. Recent studies of these community members have looked for diabetes risk, as well as other health issues. These studies are known as the C8 Health Project (C8 is another term for PFOA) (see Frisbee et al. 2009 for a description of study design).

A cross-sectional study of the exposed Mid-Ohio Valley community members did not find an association between PFOA and type 2 diabetes or fasting glucose levels (MacNeil et al. 2009), nor did a long-term study of this population (Karnes et al. 2014). A more detailed study found that those with diabetes (especially type 1) had lower levels of PFCs than those without diabetes (Conway et al. 2016). Early-life PFOA levels were not associated with obesity or overweight in adulthood (Barry et al. 2014). However they did find associations between PFOA levels and high cholesterol (Winquist and Steenland 2014). For an article describing this findins, see PFOA and High Cholesterol: Basis for the Finding of a Probable Link, published in Environmental Health Perspectives (Betts 2014).

A long-term study from Cincinnati looked at associations in the offspring of women living downstream from a PFC manufacturing plant. They found that higher maternal PFOA levels were associated with higher weight and waist circumference in their children, as well as greater BMI gains from ages 2-8 (Braun et al. 2016). Another long-term study from the Cincinnati area found PFC levels were not associated with BMI, but were associated with altered kidney and thyroid function (Blake et al. 2018). 

The World Trade Center

Large amounts of various chemical contaminants, including PFCs, were released at the time of the World Trade Center disaster in 2011. In children and adolescents who were exposed to the contaminants, those with higher levels of PFOAs had increased triglycerides, total. cholesterol, and LDL cholesterol. Perfluorohexanesulfonic acid levels, however, were associated with lower insulin resistance (Koshy et al. 2017).

Gestational Diabetes

Preconception levels of PFOA were associated with gestational diabetes in a prospective U.S. study of women with background levels of PFC exposure. Six other PFCs were also associated with an increased risk, although not statistically significant (Zhang et al. 2015). In Canadian women, first-trimester levels of most PFCs were not associated with gestational diabetes. However, this study found a higher risk of impaired glucose tolerance during pregnancy in women with moderate (second-quartile) levels of perfluorohexane sulfonate (PFHxS) (Shapiro et al. 2016). The Canadian women with higher PFOS levels also had higher gestational weight gain (Ashley-Martin et al. 2016). A study of pregnant women from Spain found that PFOS and PFHxS levels were associated with impaired glucose tolerance and gestational diabetes (Matilla-Santander et al. 2017). In Danish women with a high risk of gestational diabetes, PFHxS levels were associated with increased fasting glucose, fasting insulin, and insulin resistance. PFNA levels were associated with higher fasting insulin and beta cell function. Other PFCs were not associated, and there were no associations in pregnant women who were otherwise at low risk of gestational diabetes (Jensen et al. 2018).

However, a Colorado study found that pregnant women with higher PFOA, PFNA, PFDeA, and PFHxS levels had lower blood glucose levels (Starling et al. 2017).

In Norway, pregnant women with higher levels of PFCs had higher levels of HDL cholesterol (the "good" cholesterol) and total cholesterol (not so good) (Starling et al. 2014).

A study from China found that PFOA levels were associated with higher insulin levels, higher insulin resistance, and higher blood glucose levels in pregnant women, while PFOA tended to be associated with lower glucose levels (Wang H et al. 2018). Another study from China found that while maternal PFC exposure was not associated with risk of gestational diabetes, significant positive associations were observed between exposure to specific types of PFCs and increasing blood glucose (Wang Y et al. 2018).

The Madrid Statement

In 2015, fourteen experts published the Madrid Statement on Poly- and Perfluoroalkyl Substances (PFASs), subsequently signed by 206 scientists and professionals from 40 countries (Blum et al. 2015; Birnbaum and Grandjean 2015). 

The Statement documents the scientific consensus about the environmental persistence, bioaccumulation, and potential toxicity of these substances.

Laboratory Studies

Adult male rats exposed to PFNA experienced high blood sugar by increasing the release of glucose from the liver (Fang et al. 2012). PFOA-treated mice had increased blood glucose and insulin levels, increased insulin resistance, and higher levels of leptin and adiponectin (Du et al. 2018).

PFOA exposure reduced the production of glycogen in the liver of mice, and actually increased insulin sensitivity and glucose tolerance. While these effects may appear to be beneficial, the mechanism by which they occurred may have harmful effects in the long run-- several protein levels were also affected, and these proteins are potentially involved in diabetes and liver disease (Yan et al. 2015). Other authors also found that PFOA affects liver and beta cells in animals, and impaired liver function. Specifically, PFOA caused higher insulin and LDL cholesterol levels, and reduced glucagon, glucose, and HDL cholesterol levels (Wu et al. 2017). These authors also found that PFOA exposure also decreased fasting blood glucose, raised insulin levels, increased liver enzymes, and changed lipid levels in mice (Wu et al. 2018). Other authors found that PFOA causes higher blood glucose levels in mice, lower glycogen levels in the liver, and also-- this is new-- promoting energy consumption, especially carbohydrate consumption (Zheng et al. 2017).

Mice fed high doses of PFCs show reduced body weight and lower fat mass, via reduced food intake (Shabalina et al. 2015). (While high doses of some chemicals can cause lower weight, it may be that lower doses have opposite effects). PFOS was also found to reduce some of the unhealthy effects of a high fat diet in mice (Huck et al. 2018). 

PFOS causes changes in the liver and the intestine in zebrafish, an animal model used to study the effects of toxic chemical exposures (Cui et al. 2016). PFCs also affect the deposition of triglycerides in the liver of mice (Das et al. 2017; Hui et al. 2017). In frogs, PFOA caused lipid accumulation in the liver, and higher total cholesterol and triglyceride levels (Zhang et al. 2018a).

In adult mice, PFOS exposure caused metabolic disturbances, particularly in lipid and glucose metabolism, and perturbed gut metabolism, inducing changes associated with inflammation and metabolism (Lai et al. 2018).

The PFOS replacement chemical, 6:2 chlorinated polyfluorinated ether sulfonate (6:2 Cl-PFESA), increased liver lipid accumulation, triglycerides and LDL cholesterol, and decreased HDL and total cholesterol in mice (Zhang et al. 2018b). It appears that this replacement chemical is more toxic to the liver than the one it replaces...

Cell Culture Studies

PFOA can affect beta cells in the laboratory, decreasing cell viability and increasing cell death (Suh et al. 2017). 

Laboratory studies show that PFCs may have effects on the immune system that are consistent with autoimmune diseases. Human cells exposed to PFOS had these effects at exposures at the high end of human exposure range (Midgett et al. 2015).

PFOA increases the development of fat cells, which then accumulate triglycerides (Yamamoto et al. 2015). PFOS also increases fat cell development, and fat accumulation as well (Xu et al. 2016).

PFOS essentially causes insulin resistance in cells; scientists are working to identify the exact mechanisms involved (Qiu et al. 2016).

A replacement chemical for PFOS, perfluorobutanesulfonic acid (PFBS), promotes the differentiation of pre-fat cells into fat cells and increased triglyceride levels (Qi et al. 2018).

Exposures During Development

Pregnant/mother rats were exposed to PFOS, to see the effects on the offspring (exposed in the womb and while nursing). The offspring had low body weight from birth to weaning, and had impaired glucose tolerance, and higher insulin levels, resembling pre-diabetes (Lv et al. 2013). A study of mice showed that when pregnant mice were exposed to low doses of PFOA, their offspring had metabolic effects that differed by sex (van Esterik et al. 2016). 

Mice exposed to low doses of PFOA in the womb had reduced body weight at birth followed by excess body weight at mid-life, as well as higher insulin levels at mid-life. There was no effect of PFOA on these parameters from exposure during adulthood, showing that developmental exposures may be most critical (Hines et al. 2009).

When mother mice were exposed to PFOS during pregnancy, they had higher insulin resistance than untreated controls, suggesting a gestational-diabetes-like pattern. Early in life, their male offspring had higher insulin levels, although the female offspring had normal levels. Later in life, as adults, both groups of offspring had higher fasting glucose and insulin levels than controls. The pups fed a high-fat diet showed even greater effects than those fed a normal diet (Wan et al. 2014).

PFOS exposure in the womb appears to affect oxidative stress more in the fetus than the pregnant mother, and this could affect fetal development (Lee et al. 2015).

Studies in zebrafish show that PFOS exposure during development also affects the development of the pancreas in ways that may predispose to diabetes (Sant et al. 2016Sant et al. 2017). Also in fish, PFOA exposure during development affects glucose levels in offspring, possibly in future generations as well (Lee et al. 2017).

Perfluorobutanesulfonic acid (PFBS), a replacement chemical for PFOS, disrupts the development of the pancreas and energy homeostasis in zebrafish (Sant et al. 2018).

Diabetes Management and Complications

Some cross-sectional human studies show associations between PFCs and heart disease/cardiovascular disease (e.g., Huang et al. 2018). One longitudinal study from Sweden found no associations for seven of eight PFCs measured and heart disease (Mattsson et al. 2015), while another long-term Swedish study found PFC levels associated with a heart disease risk factor (Lind et al. 2018). Another study finds an association between PFC levels and reduced kidney function in healthy adolescents (Kataria et al. 2015). In China, PFCs are associated with high blood pressure (Bao et al. 2017). In Sweden, PFCs are associated with other signs of cardiovascular disease (Lind et al. 2017), and a long-term study found that "normal" PFC levels are associated with changes to liver function (Salihovic et al. 2018).

I have not yet seen any studies of complications and PFCs in people with diabetes. However, a laboratory study exposed rats with (type 1) diabetes to PFCs and found that the exposure caused the accumulation of triglycerides and total cholesterol in the liver, not a good thing (Fang et al. 2015). Chicken embryos exposed to PFOS, especially at the lowest doses, affect genes that control fatty acid metabolism in the liver, also not a good thing (Jacobsen et al. 2018).

The Bottom Line

In a combined analysis of the human and animal evidence, "developmental exposure to PFOA adversely affects human health based on sufficient evidence of decreased fetal growth in both human and non-human mammalian species" (Lam et al. 2014). That is, there is evidence that PFOA exposure in the womb reduces the growth of the fetus. Whether there are other related effects later in life is not yet clear. A review of the human evidence on developmental exposure to PFCs finds that "epidemiological findings are consistent and suggest possible associations with fetal and postnatal growth and immune function" (Liew et al. 2018).

PFCs are only beginning to be studied in relation to diabetes. Results may vary by exposure level, timing of exposure, and other factors.


To download or see a list of all the references cited on this page, see the collection Perfluorinated compounds and diabetes/obesity in PubMed.